This set of Koans will provide practical guidance for getting to grips with graph data structures and operations using the Neo4j open source graph database. It's part of a more comprehensive tutorial presented by the authors and others at conferences and tutorials. In fact anyone can take these materials freely and run their own tutorials.
What are Koans?
The Koan idea was borrowed from the Ruby Koans which provide a number of broken unit tests, and in fixing those tests increasingly advanced facets of Ruby are explored. The Koan model provides very rapid feedback and a structured learning path wrapped in a pre-configured environment that gets us up and running very quickly. These are very desirable characteristics when it comes to learning Neo4j, and so these Koans have adopted the same model - there are a set of (broken) unit tests, and in fixing each of them we learn some aspect of using Neo4j. As we work forwards through the Koans we'll learn more sophisticated APIs, query languages and techniques and by the end of the Koans we'll feel supremely confident about using Neo4j in production.
You'll need to be familiar with the Java programming language, and it'd be helpful if you understood unit testing too. If you like a particular IDE like Eclipse or IntelliJ, that's fine but you can run these Koans from the command line if that's the way you're wired. All the graph database knowledge you'll needed will be developed as part of completing the Koans.
Setting up Koans
The first step in setting up the koans is to clone this repository from github, or to download the latest tarball. To clone from github:
git clone firstname.lastname@example.org:jimwebber/neo4j-tutorial.git
To download the latest version of the koans as a zipfile, click on the following URI:
Once you have cloned or downloaded and unzipped the koans, you're almost ready. You just need the Neo4j binary dependencies, which you can get by running the default Apache Ant target from the Koan root directory:
We use Apache Ivy for dependency resolution and for all the many wonderful benefits Ant and Ivy provide, speed isn't one of them so you might want to go fetch a cup of tea while you're waiting for (seemingly) most of the Internet to download.
Once the default Ant target has completed, you'll find it has dropped a set of libraries in the lib directory immediately under your Koan directory. It has also run the Koans and the unit tests that validate that everything's ready and will have left a report under:
Since the Koans come complete with the answers, if you're going to benefit from following them, you'll need to delete those answers. Fortunately we've provided a sed script that does this for you. From the root directory of your Koan download, run:
Sorry to Windows users, you'll have to run this with http://www.cygwin.com/ for now.
The sed script removes all lines between
in the Koans, leaving them compiling but failing.
I want my IDE
If you'd like some help in setting up the Koans in your IDE, the Ant script contains a target called generate.eclipse.project. By issuing the command:
an Eclipse project will be created in the Koan directory. This can be opened readily with Eclipse or Intellij.
Where are my Koans?
The Koans themselves are in:
and you'll see that they're numerically increasing in number. Start with Koan01.java and run it. If you're running the Koans in your IDE, then you can run Koan01.java as a unit test and verify that a Neo4j database can be created and populated with Doctor Who data, and that you have your IDE project set up properly.
Now that you're at this point, you're ready to move to Koan02.java and start hacking on Neo4j proper. Don't forget that in the presentation directory you'll find a full set of slides that will help you navigate the Koans, and build up a picture of the Doctor Who universe.
This tutorial is presented around the world. If you're interested in participating in a class, then it'll be taught at:
Quarterly at Skillsmatter London
Once each in Melbourne, Brisbane, and Sydney at the Yow! Australia 2012 conference
Peter Neubauer, @peterneubauer